Measuring Science and Innovation for Sustainable Growth

Report

9 October 2025

2. Science and new technology development for energy and the environment

Copy link to 2. Science and new technology development for energy and the environment

Abstract

This chapter presents a comprehensive range of indicators mapping a diverse set of channels through which the generation of scientific and technological knowledge on energy and the environment may contribute to sustainable growth. It provides international comparisons, trends, and structural analysis through indicators that depict those channels. In addition to measures of scientific output, collaboration, patents and research and development, this chapter includes newly developed measures of scientific output relevant to sustainable energy and environmental goals, contributions to the Intergovernmental Panel on Climate Change reports, and science underpinning patented innovations on climate change related technologies.

In brief

Copy link to In brief

This chapter seeks answers to questions such as whether the rate of production of scientific knowledge corresponds to the scale and urgency of energy and environmental challenges or what are the characteristics of science that supports environmental and energy policy objectives and the researchers who produce it. Through existing and newly developed indicators, which provide evidence of major trends and structural issues and enable the comparison of countries' scientific and technological capabilities and outcomes, this chapter finds that:

Several recent trends in global scientific publishing are particularly marked in energy research. OECD countries’ overall scientific output is relatively unspecialised in the area of energy research, and their citation performance is appreciably lower than the world average. In contrast, the People’s Republic of China (hereafter “China”) is by a significant margin the largest contributor to scientific publication output in energy journals. Its scientific output is relatively specialised in this domain, and its rate of highly cited publications is similar to the world average.
Articles published in specialised scientific energy and environment journals are only a subset of all science relevant to energy and environmental goals. Applying a new AI-based classification tool indicating the societal goal relevance individual scientific publications reveals that the share of scientific publications that contribute to energy and environmental goals is close to 28%.
When it comes to all energy and environmentally relevant publications (not only those in specialised journals) and focusing on the 10% most highly cited scientific publications worldwide, China accounts for 40% of the world’s scientific output publications (up from 15% in 2012), followed by the European Union and the United States, with 15% and 10%, and down from 27% and 23%, respectively. Across the OECD, with few exceptions, most countries experienced a drop in the share of publications relevant to energy and environmental goals between 2012 and 2022.
There are few available indicators of science and innovation-related human resources in this area but those that are available offer some initial insights:
- In European countries, enrollment in doctoral education in formal environmental fields is rather low, at close to 1%. However, dissertation work relevant to energy and environmental goals might be significantly higher, at close to 19%, based on detailed data available for France.
- The highest shares of researchers who claim that their work is most relevant to Sustainable Development Goals (SDGs) clustered under the “planet” umbrella1 are found in agricultural and veterinary sciences (49%), natural sciences (28%) and engineering and technology (19%). The highest shares of researchers who consider their work relevant to SDG 7 (Affordable and clean energy) can be found in engineering and technology (16%) and natural sciences (5%).
- At 16%, start-ups in the energy, resources and sustainability sector are the second-most likely to be founded by PhD holders, only after health and biotechnology start-ups (25%), the paradigmatic model of academic entrepreneurship.
The Intergovernmental Panel on Climate Change (IPCC) Assessment Reports are an example of the key role played by science in building broad societal awareness of environmental challenges, their causes and potential mitigation options. The reports show increasing reliance on science produced across a broader group of countries and a particularly high rate of international scientific collaboration.
International scientific collaboration is higher for scientific publications related to energy and environmental goals than in other fields. From this higher base, collaboration has been increasing in line with broader trends for all scientific publications.
Patent data provide an indication of technological development suitable for addressing environmental challenges:
- Between 2010 and 2020, environmental patents filed in at least two intellectual property (IP) offices, one of which is a top 5 IP office, (IP5, an indicator of high-quality patenting) by inventors based in China experienced an extraordinary 604.97% increase, surpassing the United States (15.84% increase) and closely approaching the European Union (12.04%)).
- Nearly 25% of all Patent Cooperation Treaty (PCT) patent filings in China are related to environmental outcomes, compared to only 10% in the United States. The two countries had similar environment-related patent filing intensity a decade ago.
New inventions that address environmental challenges build on science to a greater extent than those based on existing high-carbon technologies. Although low-carbon technology patents rely only marginally more on non-patent literature, they account for nearly six times as many publications in their citations than high-carbon patents and those with ambiguous carbon-emission impacts. Those rely relatively more on trade literature. Specifically:
- Nearly 40% of scientific publications cited in low-carbon patents are by US-based authors, followed by China with 13% and Germany and Japan with 8% each. This distribution is expected to change, as cited publications by authors based in China are, on average, four years more recent than the average.
- In addition to engineering (16%), the key most-cited fields by low-carbon patents are chemistry (15%) and materials science (12%). As a share of all citations, computer science (7%) is 5 percentage points more important for low-carbon than high-carbon patents.
The share of venture capital for environment-related start-ups continues to grow in the European Union and China but stagnates across the OECD area.
Research and development (R&D) in the business sector contributing to energy applications is significantly higher than R&D conducted by energy and other utilities, which account, on average, for less than 1% of the total. In the United States, nearly 5% of business R&D is oriented to energy applications and 2% to environmental protection. The information and communication technology (ICT) and transport equipment industries are significant contributors to energy R&D.
According to publicly disclosed accounts, companies in the alternative energy and electricity sectors have seen robust growth in R&D expenditure since 2017, second only to the software, computer and electronics sector.

Science and its contribution to the energy and green transition

Copy link to Science and its contribution to the energy and green transition

Measurement rationale

Understanding and responding to complex and global environmental challenges like climate change, pollution or biodiversity loss largely depends on our ability to generate scientific knowledge (World Meteorological Organisation, 2023[1]). Scientific research helps build consensus on the state of planetary ecosystems, its key drivers and the scenarios for future development and human intervention. This knowledge can also inform governments and citizens with a balanced assessment of the potential impact of environmental issues (Rogelj, 2023[2]). Scientific research also plays a critical role in pushing the boundaries of knowledge that applied researchers, engineers, and designers can draw upon to develop new, viable technological solutions (Perrons, Jaffe and Le, 2020[3]). By enhancing our understanding of fundamental principles and charting unexplored pathways, basic science provides the means to develop new approaches for tackling otherwise intractable technical and socio-economic problems using current technology. Without such advances, social and political willingness to embrace environmental objectives and act on them would be severely diminished.

As discussed in Chapter 1, measuring science and research relevant to environmental sustainability and effective use of natural resources is not a trivial endeavour. Demonstrating relevance can be more straightforward for applied research, which is oriented towards specific, practical goals. However, it takes considerable effort to trace how fundamental research into principles and facts of nature and society can be relevant for future technologies and policy decisions. This chapter explores multiple channels through which these knowledge generating activities operate and provides a range of key indicators, drawing on a diverse range of data sources.

Indicators of scientific activity relating to energy and the environment

A first step in assessing the relevance of scientific output to societal goals related to energy and the environment is to focus on the academic journals that specialise in those areas. As detailed in Box 2.1, the fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy within the All Science Journal Classification (ASJC) are most explicitly connected with energy and environmental sustainability. Bibliometric analysis shows that in 2023 China was the largest country contributor to scientific publishing in these areas, surpassing the European Union and fast approaching the combined value of OECD countries, followed by the United States, India and Russia (Figure 2.1).

Box 2.1. Bibliometrics and scientific journals in energy and environment fields

Copy link to Box 2.1. Bibliometrics and scientific journals in energy and environment fields

The scientific peer review system and the body of scholarly publications it generates help provide the basis on which bibliometrics can be applied to the study of scientific research. Quantitative studies on research publications can draw on the contents and information contained in publications – which meet defined review and publication criteria – to analyse, under some assumptions, multiple dimensions of scientific production and dissemination. Specifically:

Indexed information on scientific documents helps investigate sources of scientific knowledge through the identity and affiliation of authors and the references contained in a document.
Scientific collaboration can be potentially gauged by the extent of co-authorship and/or the engagement of multiple institutions.
The relevance of the research to the broader scientific community may be inferred, in part, from the extent a publication is cited by other documents or the visibility of the title, e.g. the journal, in which a given document is published, based on its past citation record.

The interpretation of bibliometric analysis is contingent on a series of norms and incentives that vary across sectors and knowledge domains. It can also evolve over time. For example, not all scientific discoveries and research results are published in a well-defined list of international scientific journals where they can be read and cited by other researchers and a possibly wider user community. Like other forms of administrative data sources, bibliometric data do not exist to serve statistical purposes. While this does not invalidate their relevance for the statistical analysis of science, it is important to remember this key feature and apply a degree of caution when designing and interpreting bibliometric indicators.

Inferring relevance from journal classification

The ASJC system is used in the Scopus database, a bibliographic data source used by the OECD, to classify journals and conference proceedings under broad subject areas that are further divided into groups and subfields. The ASJC fields of agricultural and biological sciences; environmental science; earth and planetary science; and energy are explicitly connected with energy and environmental sustainability. Publications in journals classified in these fields can be presumed potentially relevant to sustainable growth. This approach, however, misses out on relevant scientific publications in other journals, not only those of a multidisciplinary nature. The contribution of basic science, for instance, would not necessarily be captured using this approach, thus calling for other approaches presented elsewhere in this publication.

Source: Authors, partly based on OECD and SCImago Research Group (CSIC) (2016[4]), Compendium of Bibliometric Science Indicators, https://web-archive.oecd.org/2016-09-29/415063-Bibliometrics-Compendium.pdf.

Figure 2.1. Scientific publication volume in energy and environment-themed journals, 2023
Copy link to Figure 2.1. Scientific publication volume in energy and environment-themed journals, 2023

Main contributing economies to journals in agricultural and biological sciences, environmental, earth and planetary sciences and energy fields

As a share of total domestic publication output (Figure 2.2), the rate of scientific publishing in energy and environmental fields is highest in Costa Rica (34%) and Argentina (26%). This reflects, in particular, the relative importance of agriculture for these economies and the extent to which their scientific production is oriented towards meeting economic and social needs.

Figure 2.2. Scientific publication intensity in energy and environment-themed journals, 2023
Copy link to Figure 2.2. Scientific publication intensity in energy and environment-themed journals, 2023

Economies with the largest shares of scientific publications in agricultural and biological sciences, environmental, earth sciences and energy journals

Examining trends in environmental and energy-related journal fields over the 2009-23 period among selected economies (Figure 2.3) shows that scientific publishing is only consistently on the rise as a proportion of total publishing in the area of environmental science, in what seems to reflect a growing preoccupation with the implications of natural resource utilisation, pollution and climate change. The share of agricultural and biological sciences has declined in the European Union, the United States and across OECD countries, while it has seen robust growth in China. Among those presented, China is the only major economy where the relative importance of all these fields, including energy, has increased over the period under examination.

China is by a significant margin the largest contributor to scientific publication output in energy journals, a field in which it is relatively specialised and has a rate of highly cited publications approximately similar to the world average. In contrast, the EU’s overall scientific output is relatively unspecialised in energy, and its citation performance is appreciably below the world average (Figure 2.4). European Commission’s analysis of research specialisation in each of the ‘Horizon 2020 Societal Grand Challenges’ finds that, overall, the EU is more specialised in publications related to health and less specialised in publications on secure societies and energy. In terms of the climate action, environment, resource efficiency and raw materials challenge, EU’s specialisation is around the global average, and remains relatively unchanged since 2000 (European Commission, 2024[6]).

Figure 2.3. Scientific publishing trends in the fields of energy and environment, selected economies, 2009-23
Copy link to Figure 2.3. Scientific publishing trends in the fields of energy and environment, selected economies, 2009-23

Share of domestic scientific publications in agricultural and biological sciences, environmental, earth and planetary sciences and energy journals

Figure 2.4. Specialisation and citation impact in scientific publications in the energy field, 2023
Copy link to Figure 2.4. Specialisation and citation impact in scientific publications in the energy field, 2023

Total publications displayed as bubble size

Self-reported measures of the relevance of science to energy and environmental goals

Given the limitations of determining which publications are relevant to environmental sustainability based on ASJC fields, there is a need for alternative approaches. One alternative is to ask the researchers themselves about the societal goals with which their research aligns the most. The OECD International Survey of Science (ISSA) conducted in 2021 collected data from 3 091 scientific researchers, who were asked a range of questions on their working conditions, society engagement, the impact of the coronavirus (COVID‑19) pandemic on their work, career prospects, and the alignment of their research with the United Nations (UN) SDGs. Some 94% of respondents (2 911 individuals) answered the specific question concerning the relevance of their research to one or more SDGs. Only 8% reported that their work had no relevance to the SDGs. While not necessarily representative of the entire research workforce, the survey reveals new insights regarding the orientation of science fields to various societal goals, as defined by the Sustainable Development Agenda, as well as about various characteristics of the researchers who engage in research supporting environmental sustainability and energy (Figure 2.5).

Figure 2.5. Distribution of self-reported most relevant SDG to their research among ISSA 2021 respondents
Copy link to Figure 2.5. Distribution of self-reported most relevant SDG to their research among ISSA 2021 respondents

Distribution of SDGs reported as most relevant, clustered by group, as percentage of respondents in each field

As might be expected, the highest shares of researchers who claim that their work is most relevant to the SDGs clustered under the “planet” umbrella are found in agricultural and veterinary sciences (49%), natural sciences (28%) and engineering and technology (19%). The highest shares of researchers who consider their work relevant to SDG 7 (Affordable and clean energy) can be found in engineering and technology (16%) and natural sciences (5%). According to the survey, other fields of science only have a negligible energy share.

Basic research has long been recognised as the “pacemaker of technological progress” (Bush, 1945[8]). By its very definition, the eventual application of basic research is often difficult to predict, but it is known to play a fundamental role in “clean tech” patented technologies, especially in the “deep tech” subset (Dalla Fontana and Nanda, 2023[9]). For example, advances in solid-state physics and semiconductor research in the mid-20th century paved the way for the discovery of the photovoltaic effect in silicon, which underpins today’s solar photovoltaic industry (Chapin, Fuller and Pearson, 1954[10]; Green, 2000[11]). Similarly, the development of atmospheric physics and early computational models in the 1960s laid the groundwork for modern climate modelling, which has become essential for understanding anthropogenic climate change and guiding mitigation and adaptation strategies (IPCC, 2023[12]).

These examples illustrate how knowledge generated without immediate commercial or environmental objectives can yield transformative spillovers, accelerating the shift towards low-carbon economies. According to the OECD International Survey of Science, only 21% of the researchers whose work is relevant to the “planet” SDG category report that it is purely basic research. However, the combined shares of basic research and basic and applied research are some of the highest in both the “planet” and the “prosperity – energy” categories, suggesting that basic research, including in combination with applied research, plays an important role in advancing these particular societal goals (Figure 2.6).

Figure 2.6. Orientation of research according to SDG goal relevance
Copy link to Figure 2.6. Orientation of research according to SDG goal relevance

Distribution of modalities of research orientation according to the most relevant SDG goal

New indicators of energy and “green” science

Available classification methods have several limitations, which constraints their use as a basis for indicators relevant to societal goals. While the use of “key terms” that relate to those goals to query the titles and abstracts of scientific publications may offer some useful insights into relevance of scientific publications to certain topics (e.g. Box 2.2), this approach often understates the contribution of basic science to attaining those goals.

Given these limitations, the OECD has developed an SDG classification methodology that makes it possible to analyse energy and environmental relevance of the full body of scientific publications regardless of the type of journals documents are published in and accounting for the subtle distinction between what research is about and its potential relevance and applicability. As summarised in Box 2.3, the new methodology brings together the potential of surveys, bibliographic data, and artificial intelligence (AI) tools for analysing descriptions of research work.

Box 2.2. Measurement case study: Analysis of space-based “green” research by the Danish Ministry of Higher Education and Science

Copy link to Box 2.2. Measurement case study: Analysis of space-based “green” research by the Danish Ministry of Higher Education and Science

Policymakers are often interested in how two different domains interact and assessing how competencies in one particular scientific domain contribute to advances in environmental science. The Danish Ministry of Higher Education and Science (UFM) conducted a bibliometric analysis of Denmark's space-based “green” research (Ministry of Higher Education and Science, 2022[13]), assessing its role in the country's broader green transition efforts. The report analysed the volume of relevant scientific publications, distribution across disciplines, specialisation, collaboration patterns and impact.

The study identified Denmark’s space-based green research by combining two search strategies implemented on a commercial database to identify research output that integrates space technology with environmental science. The “green research” identification follows the Ministry’s framework, covering sustainability, climate change, biodiversity, circular economy and renewable energy. The “space-based” query uses relevant keywords such as "satellite," "Earth observation," "remote sensing," and many others. The search used information on titles, abstracts and author keywords. Both the space keywords and the publications retrieved by the overall search strategy were subject to quality assurance by eight Danish universities and the Danish Meteorological Institute.

Source: Authors, based on Ministry of Education and Research (2022[13]), Rumbaseret grøn forskning – Bibliometrisk analyse af Danmarks rumbaserede grønne forskning September 2022, www.ufm.dk.

Applying the SDG classification reveals that the share of scientific publications that contributes to energy and environmental goals remained close to approximately 28% globally between 2008 and 2018 but is not uniquely concentrated in a few fields (Figure 2.7). The SDG-based classification makes it possible to revisit the classification of scientific publications by ASJC field and to examine the degree to which different fields are relevant to natural resource management and environmental sustainability. The shares of relevant publications are indeed the highest within the initially selected ASJC fields, namely energy, environmental sciences, agricultural and biological sciences and earth and planetary sciences. However, the results show that considerable scientific output within other fields was being ignored, particularly in chemical engineering and chemistry journals, which display a share of energy- and environment-relevant output of over 40%. Contributions of less than 10% are only found in the different health and medical science fields.

Box 2.3. Introducing the new OECD SDG research relevance classifier

Copy link to Box 2.3. Introducing the new OECD SDG research relevance classifier

Understanding the relevance of research to a societal goal is an important but rather different task than examining the subject or field of research. The text contained in scientific publication abstracts and R&D project descriptions lends itself to analysing subject matter in ways that significantly improve analysis that relies on journal classifications. However, it does not necessarily provide a good indication of what research is relevant for. Machine learning classifiers typically rely on topic similarity of text-based descriptions of the work but often lack explicit input into the classification process from those in a position to assess the relevance to goals such as the SDGs. This can result in understating the relevance of basic research or missing out on how goals can depend on each other.

In response to this challenge, the OECD has developed an SDG classifier using information available in the responses to the ISSA survey conducted in 2021, whose results are presented earlier in this chapter. Self-assessments of relevance have been used to train a classifier algorithm. A subset of the respondents, 69% (2 014 individuals), were successfully matched to their author profiles in Scopus 2024. A total of 11 719 publications from 2000-21 were retrieved from Scopus and associated with these authors. Each publication was labelled with its relevant SDG based on the individual respondent's ISSA survey response.

The SDG classifier relies on the SciBERT large language model, which has been trained on scientific texts, making it more adept at understanding specialised terminology. It allows for multi-label classification, reflecting the fact that the SDGs are not mutually exclusive. In multi-label classification, each label's probability is independent of the other. Since the survey respondents were able to select the most relevant SDG, including the possibility of selecting none, normalisation is performed to ensure the distribution of probabilities adds up to 100% across all possible outcomes. The thus trained classifier can be used for the SDG classification of a wide variety of texts, including scientific publications, thesis abstracts and R&D funding awards from the OECD Fundstat database.

This publication focuses on a selected group of SDGs that refer to energy (SDG 7: Clean and affordable energy) and the environmental sustainability “planet” cluster (SDG 6: Clean water and sanitation; SDG 12: Responsible consumption and production; SDG 13: Climate action; SDG 14: Life below water; SDG 15: Life on land) (hereafter referred to as “environmental”) goals. The classifier assigns probabilities to each object (publications and PhD theses in this chapter, R&D awards in Chapter 4) and uses them for fractional counting and tagging. A publication is thus tagged as relevant to energy and environmental goals if the predicted SDG cluster probability is highest for the combination of SDG 7 (Affordable and clean energy) and the SDG “planet” cluster. These are described as “energy and environmental” SDG-related publications for brevity.

Source: Authors, based on Aristodemou, L., et al. (forthcoming[14]), “Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods”.

Figure 2.7. Scientific publications relevant to energy and environmental goals, by field of journal, 2022
Copy link to Figure 2.7. Scientific publications relevant to energy and environmental goals, by field of journal, 2022

As a percentage of all publications within each journal ASJC field domain

Figure 2.8. Distribution of global scientific publications by relevance to societal goals, 2022
Copy link to Figure 2.8. Distribution of global scientific publications by relevance to societal goals, 2022

Figure 2.8 shows the estimated distribution of world’s total scientific publication output for 2022 across the different SDGs, including the “no SDG relevance” case. The energy and environment SDGs account for approximately 28% of scientific production. Among those, SDG 7 (Affordable and clean energy) is the most contributed SDG at nearly 7% of the total, followed by SDG 13 (Climate action). The results indicate that, contrary to previous work, the vast majority of scientific publication output (94%) is relevant to at least one SDG and, with exceptions such as SDG 3 (Good health and well-being) and SDG 9 (Industry, innovation and infrastructure), science efforts are broadly spread across SDGs, which in many cases overlap with each other.

The number of scientific publications deemed most likely to contribute to the energy and environmental sustainability transition increased by almost 100% from 2008 to 2022, from 500 000 to almost 1 million, thereby growing slightly over the rate of all scientific publications (Figure 2.9). However, through the COVID-19 crisis period, the share of publications relevant to energy and environmental goals declined moderately

Figure 2.9. Trend in scientific publications relevant to energy and environmental goals, 2008-22
Copy link to Figure 2.9. Trend in scientific publications relevant to energy and environmental goals, 2008-22

Thousands of publications, allocated on a fractional count basis

There have been major changes in the contribution of the largest global economies to energy and environment-relevant output. The United States and the European Union have seen a large decline in the share of relevant publications. The share in China has increased rapidly, and India has also seen a steady increase (Figure 2.10). This implies a reduction in the overall relative contribution of OECD economies to scientific output in this area, over and above the general scientific publication shift that has been taking place.

Within the OECD, most countries have experienced a drop in the share of publications relevant to the energy and environmental goals between 2012 and 2022, except Iceland and Türkiye (Figure 2.11. ). Latvia, Costa Rica, and Estonia exhibit some of the largest shares, with values of well over 30%. The United States, the Netherlands and Israel occupy the lowest ranks within the OECD, with a share of relevant publications between 14% and 18%. The EU average ranks higher than the OECD average.

Figure 2.10. Main contributors to scientific publications relevant to energy and environmental goals, 2008-22
Copy link to Figure 2.10. Main contributors to scientific publications relevant to energy and environmental goals, 2008-22

As a percentage of world's total publications relevant to energy and environmental goals

Figure 2.11. Scientific publications relevant to energy and environmental goals, selected economies, 2012 and 2022
Copy link to Figure 2.11. Scientific publications relevant to energy and environmental goals, selected economies, 2012 and 2022

OECD countries and the European Union, as a percentage of total publications

Distinguishing between publications oriented to the energy SDG 7 and other environmental sustainability SDGs under the planet cluster, it is evident that energy is particularly important in the case of Mexico, Japan and Korea for OECD countries, and several oil- and gas-producing countries among selected non-member economies, such as Kazakhstan, the Russian Federation (hereafter “Russia”) and Saudi Arabia (Figure 2.12). There is no case where the volume of scientific publications relevant to the energy SDG surpasses the ensemble of the other environmental SDG goals.

Figure 2.12. Contributions to scientific publications relevant to energy and environmental goals, selected economies, 2022
Copy link to Figure 2.12. Contributions to scientific publications relevant to energy and environmental goals, selected economies, 2022

As a percentage of total publications, OECD and selected economies

Focusing on the top 10% most-cited publications within their fields, China accounts for an even larger share of the world’s total than for all publications (close to 40% in 2022), followed by the European Union and the United States, with 15% and 10% respectively (Figure 2.13). This indicator further demonstrates that China does not only lead in environment-related product manufacturing and exports, but increasingly also in the creation of relevant knowledge (IFC, 2025[15]). Both China and India have substantially increased their share of relevant publications between 2012 and 2022, defying the otherwise almost universal decline among the group of countries that contribute the most to the world's top 10% cited publications. France, the United Kingdom, the United States and Germany experienced the largest declines. International collaboration has become one of the most significant features of scientific discovery and technological development in the 21st century and is a key vector of information exchange (OECD, 2017[16]). International scientific collaboration is particularly important in scientific research relevant to the energy and environmental goals compared to all other goals. This collaboration “premium” for environmental and energy-relevant research has remained significant despite the overall increase in collaboration (Figure 2.14).

Figure 2.13. Main contributors to the top 10% most cited scientific publications relevant to energy and environmental goals, 2012 and 2022
Copy link to Figure 2.13. Main contributors to the top 10% most cited scientific publications relevant to energy and environmental goals, 2012 and 2022

As a percentage of the world's top 10% most-cited energy and environment-relevant scientific publications

Figure 2.14. Trends in international scientific collaboration relevant to energy and environmental goals, 2000-2022
Copy link to Figure 2.14. Trends in international scientific collaboration relevant to energy and environmental goals, 2000-2022

As a percentage of all publications

The pattern holds across most of the surveyed countries. However, in some, the difference in collaboration rates between research relevant to environmental sustainability and energy and other research is small (Figure 2.15). As Figure 2.16 shows, international collaboration within the scientific domain relevant to environmental sustainability has increased over time in all countries but India and the Russian Federation.

Figure 2.15. International scientific collaboration relevant to energy and environmental goals, selected economies, 2022
Copy link to Figure 2.15. International scientific collaboration relevant to energy and environmental goals, selected economies, 2022

As a percentage of domestically authored documents, fractional counts, selected economies

Figure 2.16. Intensity of international collaboration in energy and environment relevant scientific publications, selected economies, 2012 and 2022
Copy link to Figure 2.16. Intensity of international collaboration in energy and environment relevant scientific publications, selected economies, 2012 and 2022

As a percentage of domestically authored documents, fractional counts

Research and innovation workforce capabilities

Copy link to Research and innovation workforce capabilities

Measurement rationale

Progress towards achieving sustainable growth is entirely dependent on the level of relevant knowledge and skills within a population. The availability and productive engagement of specialised talent underpins all possible contributions of science and innovation, as laid out in Chapter 1. There is a fast-growing literature on skills required for the environmental sustainability transition and measuring their current and expected labour market footprint. For example, across OECD countries, 20% of the workforce is estimated to be employed in “green-driven” occupations (Figure 2.17), a category that comprises not only jobs that directly contribute to emission reductions but also those that are not currently but are likely to be, in demand because they provide goods and services needed for environment-related activities (OECD, 2024[17]). Some 14% of “green-driven” employment as defined in the OECD Employment Outlook 2024 is deemed to be in new and emerging occupations that did not previously exist and, therefore, imply a degree of innovation.

Figure 2.17. Green-driven and greenhouse-gas-intensive occupations, selected economies, 2015-19
Copy link to Figure 2.17. Green-driven and greenhouse-gas-intensive occupations, selected economies, 2015-19

As a percentage of all employment, period average

The OECD Employment Outlook 2024 ([18]) notes that transitioning from greenhouse gas (GHG)-intensive to innovative, “green-driven” occupations may be significantly more challenging for low-skilled workers. Ensuring sufficient and appropriate training for low-skilled workers will be paramount to addressing both skill shortages in emerging environment-related industries and supporting the learning needs of low-skilled workers. OECD analysis suggests that most GHG-intensive occupations share similar skill requirements with non-GHG-intensive occupations, suggesting that transitions within highly polluting sectors are feasible with well-targeted reskilling.

However, less attention has been paid to monitoring the nurturing and deployment of advanced research and innovation (R&I) talent that can push scientific and technological boundaries in a direction consistent with sustainable growth objectives. This section explores a limited set of available and experimental indicators that aim to overcome the dual challenge associated with assessing the size of the R&I talent pool, a significant undertaking in its own right (Box 2.4), and obtaining evidence on its capability to contribute to environmental sustainability goals through STI activities.

Box 2.4. Concepts for measuring R&I talent and its contribution to environmental goals

Copy link to Box 2.4. Concepts for measuring R&I talent and its contribution to environmental goals

The broad concept of “R&I talent” can be approached as comprising the collectives that are effectively involved or have the potential to contribute to R&D or other related STI activities. Definitions for R&D and innovation are provided in the OECD Frascati and Oslo manuals, respectively. A broad view of R&I talent includes not only those actively involved in R&D (R&D personnel) and innovation but also individuals who, while not actively involved in such activities at any given point, have the potential to do so given their competences, such as educational qualifications, skills and experience.

With international growth in education rates, educational attainment has become an important indicator of the capacity to engage in R&I activities at any level. Although formal qualifications are neither necessary nor sufficient conditions of capability to engage in R&I activities, employers place increasing emphasis on educational accreditation standards. For example, doctoral degree holders have attained the highest level of formal qualification (International Standard Classification of Education [ISCED] 8). They have been specifically trained to conduct original research, and, as a result, as they graduate, they are uniquely well-qualified to contribute to generating scientific knowledge. It is, therefore, relevant to ask whether it is also possible to assess the contribution of these individuals to specific societal goals.

At PhD or other levels, information on the field of education for programmes, as categorised by the ISCED-F classification of fields of education (UNESCO, 2015[19]), provides an additional monitoring dimension for understanding the specific expertise and disciplinary backgrounds within the R&I workforce. Six ISCED-F classification classes are explicitly aligned with the natural environment: 1) environment; 2) environment (not defined); 3) environmental sciences; 4) natural environment and wildlife; 5) environment n.e.c. (not elsewhere classified); and 6) environmental protection technology. However, these cover a relatively minor share of all knowledge domains with proven possible relevance to energy and environmental goals.

Occupational and competence-based frameworks provide a complementary perspective, as both likely “research” and “green”-related jobs have been mapped out in separate thematic exercises. The key challenge ultimately stems from what information can be reliably retrieved from existing and yet-to-be-developed data sources. If data are only generated or made available at highly aggregated occupational levels, it might be misleading to assume that these occupations, across different countries, will have similar propensities to be both environment and R&I-related.

The Research and Innovation Careers Observatory (ReICO) aims to serve as a comprehensive access point for international statistics, analytical tools and resources focused on R&I careers. In collaboration with national governments and various stakeholders, ReICO is developing new methodologies, indicators and insights to support such goals. Working to identify the relevance of R&I talent to specific goals, such as the energy and environmental sustainability transition, will be a likely area of policy user interest in this new monitoring tool.

Source: OECD (n.d.[20]), Research and Innovation Careers Observatory, https://www.oecd.org/en/networks/research-and-innovation-careers-observatory.html.

Indicators of R&I talent development in energy- and environment-related areas

Students enrolled in environment-related doctoral programmes

Statistics on enrolment and graduation by education programmes can be used to map potential additions to the scientific research workforce relevant to sustainability, using the more explicitly connected categories in the ISCED-F classification set out in Box 2.4. Although data with this level of detail are not available for the ensemble of OECD countries, Eurostat data show that within the European Union, Spain, Italy and France had the most students enrolled in PhD-equivalent environment programmes in 2023. As a share of total doctoral students, environmental subjects account for slightly over 1% among EU and associated countries, with Switzerland and Estonia in the lead (Figure 2.19).

Figure 2.18. Students enrolled in doctoral education in environmental fields in EU countries, 2023
Copy link to Figure 2.18. Students enrolled in doctoral education in environmental fields in EU countries, 2023

Figure 2.19. Share of students in doctoral education in environmental fields in EU countries, 2023
Copy link to Figure 2.19. Share of students in doctoral education in environmental fields in EU countries, 2023

As a percentage of the total number of students enrolled across all doctoral education programmes

Experimental indicators of doctoral dissertations related to energy and environmental goals

While classifications of educational programmes provide a similar guide to measurement as journal fields do for classifying scientific output, the newly developed OECD SDG classifier presented earlier (Box 2.3) can be deployed in datasets comprising doctoral dissertation titles, abstracts and full text to determine which ones are relevant to the energy and environment-related SDGs. The application of this SDG classifier is pertinent as it operates with a similar type of research content. Given limited data availability at the international level, this section demonstrates this measurement approach using publicly available data on French doctoral dissertations submitted between 2018 and 2023. Principal component analysis (PCA), a dimensionality-reduction statistical technique with applications in exploratory data analysis and visualisation, helps map the SDG relevance space for doctoral theses in France (Figure 2.20).

Figure 2.20. Patterns of SDG relevance in French doctoral dissertations, 2018-23
Copy link to Figure 2.20. Patterns of SDG relevance in French doctoral dissertations, 2018-23

Principal component analysis (PCA) biplot of estimated SDG relevance probabilities

Across the population of PhD candidates, “planet” SDG-related dissertations account for 19%, while the energy share is 5%. The combined 24% value is close to the 22% share found for scientific publications for France. Mapped onto a two-dimensional space, the position of the variables depicting theses’ probabilities for 17 SDGs and no SDG relevance shows clusters that resemble common groupings for SDGs previously introduced but also highlight how the “zero hunger” goal is very closely related to “planet” objectives, in particular, the “Life on land” objective, showing the close connection between science environmental goals and impacts on food systems.

Measures of R&I workforce talent relevant to the energy and green transition

Researchers’ demographic characteristics

The ISSA survey discussed in the previous section also provides insights into the demographic characteristics and working conditions of scientists and researchers whose work is relevant to different SDGs. Understanding patterns can help identify potential trends of under- or over-representation of different social groups and different incentives to pursue research careers with a range of possible sustainability impacts.

Based on the non-probabilistic sample obtained from the survey (Figure 2.21), the age of researchers whose work is most relevant to the environment SDGs is found to be only marginally lower than the average, while scientists whose work is considered most relevant to SDG 7 (Clean and affordable energy) are among the youngest, almost five years below the sample average of 45.9. Younger researchers appear to be more likely to report SDG relevance other than “Quality education” SDGs; this includes respondents who engage in scholarship-driven research rather than research that is necessarily focused on education.

Figure 2.21. Demographic features of ISSA respondents according to SDG relevance
Copy link to Figure 2.21. Demographic features of ISSA respondents according to SDG relevance

There are no appreciable differences between men and women in terms of the relevance of their work to the environmental SDGs, while men appear to be twice as likely to pursue scientific research work relative to SDG 7 (Affordable and clean energy). Women are particularly more likely to contribute their scientific work to the “Health and society” SDG cluster, while men are more oriented toward the “Economic prosperity” SDG cluster.

Box 2.5. Measurement case study: beyond the R&D workforce – the UK R&I Workforce Survey

Copy link to Box 2.5. Measurement case study: beyond the R&D workforce – the UK R&I Workforce Survey

The 2022 UK R&I Workforce Survey covers the full diversity of occupations in the R&I workforce, including trainee and experienced researchers, technicians, engineers, R&I leaders and managers. All types of specialisms and sectors (private, public or non-profit) are surveyed. Among other things, the first wave of the workforce survey issued in 2022 supported the business case for the Green Future Fellowship, a programme supporting scientists, researchers and innovators to develop and scale up their breakthrough climate solutions.

While the survey’s definitions do not necessarily align with the Frascati or Oslo Manuals, the survey provides a useful insight into the R&I workforce active within the “Energy and Environment” technology family, which has been selected within the UK government’s innovation strategy as part of a larger group of seven technology classes identified as being of strategic interest. Energy and Environment technologies were the second-most commonly reported technology family in terms of relevance to the role of the surveyed workforce (22%).

The survey also details the proportion of respondents whose work relates to each of the technology families for each sector. Energy and Environment technologies were more prevalent among R&I workers in further education colleges (37%), as well as private sector businesses and public sector research organisations (both 27%) (Figure 2.22, Panel B). With 36%, the Energy and Environment technology domain also ranks among the top three in terms of share of business owners/sole traders (Figure 2.22, Panel A). This suggests a key role for start-ups in R&I in this area.

Figure 2.22. R&I workforce distribution across technology families and sectors
Copy link to Figure 2.22. R&I workforce distribution across technology families and sectors

Inventors and entrepreneurs

The potential of academic entrepreneurship to transfer valuable scientific knowledge into commercial applications (the application of knowledge is the subject of Chapter 3) is widely recognised by policymakers and economists alike. Indeed, there are various prominent examples among today's most prolific companies that were founded by scientists who came from academia, including Google, DeepMind, BioNTech and Moderna. The unique scientific foundations are frequently assumed to generate innovations with stronger breakthrough potential and high social value (Berger, Dechezleprêtre and Kirpichev Cherezov, forthcoming[24]).

Building on the definition used by Roche, Conti and Rothaermel (2020[25]), an OECD study (Berger, Dechezleprêtre and Kirpichev Cherezov, forthcoming[24]) examines start-ups founded by academic entrepreneurs (AEs), i.e. founders with a doctoral degree. The energy, resources and sustainability sector has the second-highest share of founders with doctoral degrees after the health and biotechnology sector. Start-ups in the media, real estate and travel sectors present the lowest shares of PhD-holding founders (Figure 2.23).

Figure 2.23. Start-ups with PhD-holding founders in OECD countries, by sector, 2000-22
Copy link to Figure 2.23. Start-ups with PhD-holding founders in OECD countries, by sector, 2000-22

As a percentage of all start-ups in the OECD/STI Start-ups Database

The influence of science on public understanding and policy

Copy link to The influence of science on public understanding and policy

Measurement rationale

As noted in Chapter 1, science is indispensable in building broad societal awareness of environmental challenges and their causes. They can also inform an objective and risk-based assessment of options for policy at multiple levels. For example, the accumulation of knowledge about climate change since the 1960s has enabled the scientific community to alert decision makers and the public to the risks of climate change, helping identify avenues for mitigation and their relative costs (Hallegatte et al., 2016[26]). The IPCC has been instrumental and is key in supplying scientific evidence to the United Nations Framework Convention on Climate Change (UNFCCC), an international treaty established in 1992 to co‑ordinate global efforts to combat climate change through mitigation, adaptation and financial support.

Box 2.6. Data sources and methods for the analysis of scientific influence on environmental policy

Copy link to Box 2.6. Data sources and methods for the analysis of scientific influence on environmental policy

Science informs and underpins policy at different levels and through many different, hard-to-quantify channels. As a result, policymakers across countries find it difficult to consistently measure the overall uptake of science. However, tracking citation links in systematic scientific compilations, such as in IPCC reports, offers a unique opportunity to gain insight into the nexus between science and policy in the environmental domain. IPCC citations provide an important case to demonstrate the approach by referring to a well-defined, authoritative and transparent scientific consensus mechanism.

While the citations within the latest IPCC report can be comprehensively analysed thanks to structured bibliometric data issued by the IPCC itself, no such data are available for the earlier IPCC reports. To bridge the data gap, this section draws partly on Overton (Szomszor and Adie, 2022[27]), a private data provider that developed a database linking policy documents to the documents they cite and are cited by, by web-crawling publicly accessible documents published by a curated list of over 30 000 organisations. This makes it possible to trace country contributions to Assessment Reports 1‑5 published between 1990 and 2023. Citation parsing is, however, less complete for older documents, and the coverage has been deemed insufficient for Assessment Reports 1 and 2. The database has been used to study also the relevance of climate change research to policy more broadly (Bornmann et al., 2022[28]).

It is still relatively early to make robust comparisons between national policies’ degrees of reliance on science for their environmental and related policies. Explicit recognition of scientific sources is a heterogeneous and evolving practice, dependent on the nature of the policy work, the publication of its outcomes, and transparency on the underlying processes. Scientific influence may not always leave a citation trace, while references may not reflect actual influence.

Indicators of relevance and influence in IPCC reports

The work of the IPCC is organised into three main working groups. Working Group I assesses the physical science basis of climate change, confirming human influence on global warming; Working Group II evaluates climate change impacts, vulnerabilities and adaptation strategies; and Working Group III explores mitigation solutions to reduce GHG emissions. These groups provide a scientific foundation for global climate policies and action under the UNFCCC and Paris Agreement. Findings are published in the form of Assessment Reports, which are released approximately every four years.

In terms of absolute contribution to the most recent 6^th IPCC report, based on fractional counts, the United States leads, followed by the United Kingdom, Australia, Germany and China. Countries have different relative strengths within each working group, however. France, Switzerland and Japan contribute the most, in terms of share of publications, to the physical science basis, ranging from 44% to 38%. South Africa, the Netherlands and Australia contribute the most to Working Group II (72% to 68%), while Denmark, Austria and Sweden contribute the most to Working Group III (37% to 34%) (Figure 2.24).

Figure 2.24. Main contributing countries to scientific publications cited in the 6th IPCC Assessment Report, by working group
Copy link to Figure 2.24. Main contributing countries to scientific publications cited in the 6th IPCC Assessment Report, by working group

Top 20 contributing countries, country of affiliation-based count of scientific publications cited in the IPCC report

Figure 2.25. Main contributing countries to publications cited in IPCC Assessment Reports 3 to 6, 2001-23
Copy link to Figure 2.25. Main contributing countries to publications cited in IPCC Assessment Reports 3 to 6, 2001-23

Although the top two contributing countries to the IPCC reports, the United States and the United Kingdom, have remained the same, at least since the third Assessment Report, published in 2001, the relative importance of the United States in the top ten contributing countries to each Assessment Report has declined over time. In the two recent Assessment Reports, China has made an increasing contribution, rising from 4% to 8% (Figure 2.25).

International collaboration plays an important role in the scientific publications underpinning the IPCC reports. In the 6^th Assessment Report, almost one-half of the cited scientific publications are based on single-country authorship, while 34% are based on multilateral collaboration (Figure 2.26). The rate of multilateral collaboration changes across the three working groups, with Working Group III (Mitigation options) relying more extensively on multilateral collaboration compared to Working Group I (Physical science basis) and Working Group II (Vulnerability of socio-economic and natural systems).

Figure 2.26. International collaboration in the 6^th IPCC Assessment Report, by working group
Copy link to Figure 2.26. International collaboration in the 6<sup>th</sup> IPCC Assessment Report, by working group

A broad range of scientific disciplines contributes to the work of the IPCC. In terms of the thematic distribution of the science cited in the 6^th Assessment Report, “Environmental science” and “Earth and planetary sciences” occupy the top two ranks, while “Agricultural and biological sciences” are also in the top five. “Social sciences” and “Multidisciplinary science”, however, also play an important role in the science base supporting the 6^th IPCC Assessment Report, although the role of the latter appears to be less important in terms of supporting climate change mitigation and adaptation technologies (Figure 2.27).

Figure 2.27. Main scientific fields contributing to the 6^th IPCC Assessment Report
Copy link to Figure 2.27. Main scientific fields contributing to the 6<sup>th</sup> IPCC Assessment Report

Top 20 ASJC journal fields, fractional counts

Inventive activity in energy and environmental technology

Copy link to Inventive activity in energy and environmental technology

Measurement rationale

Given the urgent need to push forward technology boundaries in order to develop environmentally superior goods and services, it is important to understand the dynamics in the technological fields that are aligned with sustainable growth as well as those associated with highly carbon-intensive economic activities. An invention is a unique or novel device, method, composition, idea or process. It can take the form of an improvement upon a machine, product or process to increase efficiency or lower costs. This section focuses on the use of patent statistics as indicators of inventions. Because patents establish claims on what their protected inventions can do and those are “signed off” by experts in patent offices, there is an explicit, objective and direct relevance link to the achievement of energy or environmental goals that is particularly valuable. This explains the significant contribution of these indicators to monitoring STI for sustainable growth (Box 2.7).

Box 2.7. Measuring inventions relevant to the energy and green transition

Copy link to Box 2.7. Measuring inventions relevant to the energy and green transition

Patents as indicators of invention

Patents are a means of protecting inventions developed by firms, institutions or individuals, and as such, they may be interpreted as indicators of invention. Before an invention can become an innovation, further entrepreneurial efforts are required to develop it and put it into use or bring it to the market. Patent indicators convey information on the output and processes of inventive activities (OECD, 2009[29]).

Patent related indicators presented in the report are based on patent applications filed under the Patent Co-operation Treaty (PCT) and on patent families:

A PCT patent application is an international filing process that streamlines entry into multiple jurisdictions.
Patent families refer to patents filed in different jurisdictions to protect a same invention. Patents that are filed in more than one Industrial Property (IP) office are often considered to be of higher quality because seeking protection in multiple major markets requires significant investment, indicating the applicant's confidence in the patent's value and broad commercial potential. In the case of this publication, IP5 patents are defined as those filed in at least two offices, one of which within the five largest IP offices (Dernis et al., 2015[30]), i.e. the European Patent Office (EPO), the Japan Patent Office (JPO), the Korean Intellectual Property Office (KIPO), the China National Intellectual Property Administration (CNIPA), and the United States Patent and Trademark Office (USPTO).

Measuring the relevance of patents to environmental objectives

Patents provide information on the technological content of the invention (notably its technical field) and the geographical location of the inventive process. The main benefit, in the context of this publication, is that the detailed technological taxonomy in which patent documents are classified, the International Patent Classification (IPC) scheme and the Cooperative Patent Classification (CPC) scheme that extends it, and a number of classification approaches that have been subsequently developed, make it possible to distinguish between innovations based on their environmental attributes.

The two most widely adopted approaches to identifying patented innovations relevant to environmental sustainability are:

the environment-related technology (ENV-TECH) classification system developed by the OECD (Haščič and Migotto, 2015[31])
the Y02/Y04S tagging scheme for Climate Change Mitigation and Adaptation Technologies implemented by the European Patent Office (EPO).

The Y02/Y04S patent tagging scheme has several key subclasses that allow for disaggregation by technology families that broadly align with key economic sectors (energy, transport, industry, agriculture, etc.). The related Y04S classification further extends to smart grid technologies and systems that enhance energy distribution and consumption efficiency. The ENV-TECH overlaps with the Y02 definition but is broader.

For monitoring and comparison purposes, the International Energy Agency (IEA) and the EPO have developed an approach to identify “high-carbon” patents (IEA/OECD, 2021[32]), which has been further extended by the OECD within the STI Microdata Lab2.

There is an additional category of “grey” patents, namely those that tend to target efficiency improvements in high-carbon or otherwise polluting technologies, for instance technologies which improve the energy efficiency of the internal combustion engine (Aghion et al., 2016[33]). “Grey” patents definition is adapted from the list of CPC class symbol codes initially published in Dechezleprêtre, Martin and Mohen (2013[34]) and Aghion et al.,(2016[33])3.

Limitations

The limitations of using patents as a proxy for innovation are widely discussed in the literature [e.g. Calel and Dechezleprêtre (2016[35]). Patent filings are not necessarily comparable indicators of inventive activity across different technology domains (a patent in one field may cover a narrower set of claims than in another), and the propensity to patent varies substantially by country and sector, which is why absolute patent counts have limited usefulness as an indicator on its own. Some also argue that patents should not be used as an indicator of R&D success (Reeb and Zhao, 2020[36]); however, due to the features outlined above, patents remain one of the most relevant indicators of R&D output and a key, yet not exclusive, precursor to innovation.

Source: OECD, based on multiple sources, including OECD (2009[29]), Patent Statistics Manual, https://doi.org/10.1787/9789264056442-en.

Patent indicators of inventions with different environmental attributes

Following a period of robust growth between 2000 and 2010, patenting of environment-related and climate-related technologies declined as a share of global patenting (Figure 2.28).

Figure 2.28. Patents in environment-related, low-carbon, grey and high-carbon technologies, 2000-2022
Copy link to Figure 2.28. Patents in environment-related, low-carbon, grey and high-carbon technologies, 2000-2022

As a percentage of total patent applications filed under the Patent Cooperation Treaty (PCT)

This is a signal of relative innovation priorities shifting elsewhere and a potential source of concern, given the urgency of the environmental sustainability transition. The shares of high-carbon and grey patents have remained relatively stable in comparison, though even these have seen a minor decline in share. The decline in the share of sustainability-related patents has been widely commented on in the academic literature (Popp et al., 2020[38]; Martin and Verhoeven, 2022[39]; Aghion et al., 2016[33]), with suggested reasons including energy prices, unfavourable policy changes, the exhaustion of the most promising innovation opportunities or the constrained availability of capital in the aftermath of the 2008 financial crisis. The breakdown of climate change mitigation patents into subcategories reveals that patenting in low-carbon energy technologies has experienced the largest decline in terms of its share of the total, followed by low-carbon technologies in transport and, to a lesser degree, buildings, industry and agriculture (Figure 2.29).

Figure 2.29. Trends in patenting for selected climate change mitigation categories, 2000-2022
Copy link to Figure 2.29. Trends in patenting for selected climate change mitigation categories, 2000-2022

As a percentage of total patent applications filed under the Patent Cooperation Treaty (PCT)

The rapid increase of patenting in environment-related technologies in China can be observed both with respect to patents filed under the PCT and with respect to IP5 patent families. With respect to IP5 patent families, China experienced an extraordinary 604.97% increase in patent filings between 2010 and 2020. Korea also showed strong expansion with a 79.06% rise, while the United States (15.84%) and EU27 (12.04%) saw moderate but steady growth. In contrast, Japan (-2.05%) and Germany (-2.19%) experienced slight declines, possibly indicating shifts in their innovation landscapes or a preference for alternative intellectual property strategies (Figure 2.30). The considerable momentum gathered by China in environment-related knowledge creation, demonstrated though patent filings as well as other indicators, deserves further assessment. Recent work by the IFC (2025[15]) shows that while Chinese environmental patents receive high citation counts, their estimated economic value, based on measurement methodology established in Guillard et al. (2021[40]) is somewhat lower compared to those from other middle- or low-income and high-income countries

Figure 2.30. Selected contributing economies to environment-related patenting, 2000-2022 (PCT filings) and 2000-2020 (IP5 patent families)
Copy link to Figure 2.30. Selected contributing economies to environment-related patenting, 2000-2022 (PCT filings) and 2000-2020 (IP5 patent families)

The share of environment-related patents ranges between 8.2% in Ireland and almost 25% in Denmark (Figure 2.31). Examining the changes in share over time, the global average remains constant; however, there have been some notable declines in share in Greece, Portugal, Costa Rica and Lithuania. However, it must be noted that given the limited PCT patent portfolios in smaller economies, these changes might be limited only to several patents.

Figure 2.31. Differences and change in environment technology patent intensity across selected countries between 2003-12 and 2013-22
Copy link to Figure 2.31. Differences and change in environment technology patent intensity across selected countries between 2003-12 and 2013-22

Share of ENV-TECH patents in total, patent applications filed under the Patent Cooperation Treaty (PCT)

Adaptation to climate change is crucial to reduce the risks posed by extreme weather events, rising sea levels, and shifting ecosystems, which threaten lives, infrastructure and economies. Societies can minimise damage and maintain stability by implementing strategies such as resilient infrastructure, sustainable water management, and climate-smart agriculture. The Y02A patent tag is part of the CPC system and covers technologies for climate change adaptation, including water supply, flood protection, resilient infrastructure, and agriculture. It helps track innovation in areas like desalination, drought-resistant crops, and buildings designed for extreme weather. With almost 5% of total patents dedicated to climate change adaptation technology, Norway is in the lead, followed by Chile and Iceland (Figure 2.32). Changes from 2003-12 to 2013-22 remain minor for most countries, although Chile, Mexico and Costa Rica have experienced substantial drops, while Norway, Colombia and Latvia have seen an increase.

The content of climate change relevant patenting varies substantially among technology domains. Electrical machinery, engines, pumps and turbines, and environmental technology comprise over 30% of climate change relevant patents, followed by materials, thermal devices, and transport. The share of patents in the engines, pumps and turbines, environmental technology, and thermal devices dropped by several percentage points between 2008-12 and 2018-22 (Figure 2.33).

Figure 2.32. Differences and change in patenting intensity for climate change adaptation technologies across selected countries between 2003-12 and 2013-22
Copy link to Figure 2.32. Differences and change in patenting intensity for climate change adaptation technologies across selected countries between 2003-12 and 2013-22

Share of Y02A patents in total, patent applications filed under the Patent Cooperation Treaty (PCT)

Figure 2.33. Share of climate change mitigation and adaptation technologies across WIPO domains, 2008-12 and 2018-22
Copy link to Figure 2.33. Share of climate change mitigation and adaptation technologies across WIPO domains, 2008-12 and 2018-22

Science for green technology

Copy link to Science for green technology

Measurement rationale

There is broad agreement on the importance of scientific knowledge as a foundation for major technological advances, and on the need for new insights into fundamental principles of matter and energy. Yet, evidence remains limited on the specific contribution of science and scientific institutions to breakthrough innovations. Given the constraints of current technologies, there is significant scope (Box 2.8) to examine how scientific research underpins inventions relevant to environmental sustainability. A clearer understanding of how energy and environmental technologies build on basic and applied science would provide valuable guidance for policymakers.

Box 2.8. Measuring the relevance of science for climate change related technology

Copy link to Box 2.8. Measuring the relevance of science for climate change related technology

Patents citing non-patent literature as traces of science's impact on technology

Prior knowledge on which patented inventions rely may encompass patents, scientific work and other sources. Information contained in references to scientific literature, conference proceedings, databases, and other relevant literature, known as non-patent literature (NPL), can shed light on the knowledge flows between science and innovation, as well as on the scientific pillars of technology. NPL referenced by inventors and patent examiners includes scientific literature as well as other literature that reflects the state of the art in an area, such as trade journals, standards, etc. Tracing which of those references link to scientific publications makes it possible to characterise the science considered relevant for patented inventions.

With some limitations, such as the fact that there may be reasons for citing science beyond the recognition of relevant prior knowledge, citations in patent documents to NPL offer an opportunity to trace the impact of scientific research (van Raan, 2017[42]) and gather valuable insights into the differences in reliance on science between patents with different environmental attributes (Perrons, Jaffe and Le, 2020[3]). Research suggests that there is a correlation between the quality of science cited in a patent and its value, with patents that ranked high on the quality of linked science being twice as valuable as patents linked to low-quality science, which, in turn, are about as valuable as patents without a direct science link (Poege et al., 2019[43]).

Data sources and methods

The indicators in this section rely on the “Reliance on science” dataset, which offers an open-access compilation of NPLs from the front pages or the text of patents granted globally to scientific papers (Marx and Fuegi, 2020[44]; 2021[45]). This dataset offers a comprehensive curated linkage between NPLs and scientific literature, encompassing a collection of approximately 16.8 million citations extracted from the full text of patents granted by the USPTO and the EPO. These 16.8 million citations initially represent a broad pool identified through automated matching algorithms, extracted from the patent’s front page, text or both, and each is assigned a confidence score that reflects the accuracy of the patent-to-article linkage. Citations with the highest confidence scores indicate very high reliability of the match. To ensure methodological robustness, recall performance has been assessed using a carefully designed validation procedure involving 5 939 randomly selected, cross-verified “known good” citations. These citations represent a high-confidence benchmark that the dataset creators have never encountered, ensuring an unbiased recall assessment. At a precision threshold of 99.4%, the recall rate is 78% for the complete evaluation set. Moreover, recall performance increases to 88% when focusing specifically on patent citations originally specified without errors or inconsistencies. The sample was constrained to citations listed explicitly on the front pages of patents, and selecting only those matches with a confidence score higher than five resulted in a refined subset. This filtering reduced the dataset from the original 16.8 million citations to 5.7 million records, of which 98.2% successfully matched with extracted complete records from the OpenAlex database. With that connection, further information can be extracted about the scientific publications cited and their authors’ affiliations, either from within OpenAlex or other bibliographic databases linked through the publication’s digital object identifiers (DOIs).

The process of linking patents to scientific literature is fraught with challenges, however, primarily due to the unstructured nature of patent citations, which often contain inconsistencies such as misspellings or incomplete information. While not all NPL is peer-reviewed science, a significant share is, and therefore, this component can be considered suggestive of potential spillovers from science to private and public R&D efforts. For all technologies, the trend is towards greater reliance on NPL in patented innovations over time, although this can also reflect a change in examination practices. Given its high potential, work on this newly matched database is ongoing as a joint endeavour of the OECD Committees for Scientific and Technological Policy (CSTP) and Industry Innovation and Entrepreneurship (CIIE).

Green inventions’ reliance on knowledge sources other than patents

Climate change mitigation and adaptation patents rely on non-patent literature (NPL) significantly more than high-carbon patents and “grey” patents (i.e. those that improve the efficiency of high-carbon technologies) (Figure 2.34). This result is also echoed in the literature, albeit for a smaller sample of patents constrained to energy (Persoon, Bekkers and Alkemade, 2020[46]).

Figure 2.34. Patents citing non-patent literature, by climate change related attributes, 2007-22
Copy link to Figure 2.34. Patents citing non-patent literature, by climate change related attributes, 2007-22

As a percentage of all patents in each patent technology category

Climate change mitigation and adaptation patents tend to have a higher intensity of citation of NPL compared to high-carbon and grey patents across most countries, except Denmark (Figure 2.35). It is important to note, however, that patents relating to climate change (positively or negatively) are overall less likely to cite the NPL than the average patented invention, a body that is dominated in most countries by ICT and pharmaceutical patents. It is not particularly meaningful to compare such very different technology domains, but it is more informative to focus on comparing within domains with an energy technology orientation but with different environmental implications, as discussed in the definitions provided in Box 2.7.

Figure 2.35. Patents citing non-patent literature, by climate change related attributes, in selected economies, 2017-21
Copy link to Figure 2.35. Patents citing non-patent literature, by climate change related attributes, in selected economies, 2017-21

As a percentage of all domestic patents, by technology and inventor’s location, selected economies

Reliance on NPL in climate change mitigation and adaptation patents increased across most countries between 2007-11 and 2017-21 (Figure 2.36). This might suggest an increasing shift towards “deep tech” technologies within this subset, which could be expected to be more dependent on science. The possibility of the “low-hanging fruit” already being picked and marginal low-carbon innovation requiring more effort (and possibly deeper scientific foundations) has been raised as a possible reason for the decline in low-carbon patent share observed in Figure 2.29 (Popp et al., 2020[38]).

Figure 2.36. Climate change mitigation and adaptation patents citing non-patent literature, in selected economies, 2007-11 and 2017-21
Copy link to Figure 2.36. Climate change mitigation and adaptation patents citing non-patent literature, in selected economies, 2007-11 and 2017-21

As a percentage of each economy’s climate change mitigation and adaptation patents, by inventor’s location

Scientific foundations of climate change relevant inventions

Climate change mitigation and adaptation technological innovation is driven by diverse scientific disciplines, often emerging at the intersection of multiple fields. New OECD analysis shows that while there are similarities between low-carbon, high-carbon and grey (i.e. ambiguous) technology patents (see Box 2.7 for definitions) with respect to the scientific domains they tend to cite, there are some important differences.

Inventions on technologies that help address environmental challenges build on science to a greater extent than inventions enhancing existing polluting technologies. Although new low-carbon technology patents rely only marginally more on NPL, as shown in the previous subsection, they account for nearly six times as many publications in their citations as high-carbon patents and those with ambiguous carbon-emission impacts. The implication is that the latter two rely relatively more on trade literature, a possible indication of differences in technology maturity.

In addition to engineering (17%), the key fields most cited by low-carbon patents are chemistry (15%) and materials science (12%). As a share of all citations, computer science (7%) is 5 percentage points more important for low-carbon than for high-carbon patents (2%) (Figure 2.37). The wide-ranging nature of these scientific influences underscores the challenge of pinpointing a single dominant field driving the development of new low-carbon technologies and associated innovations.

Examining the contribution of individual countries to the scientific basis that underpins climate change-related patents reveals that nearly 40% of scientific publications cited in low-carbon patents are by US-based authors, followed by China with 13% and Germany and Japan with 8% each. This distribution is expected to change as cited publications by authors based in China are four years more recent than the average (Figure 2.38).

Figure 2.37. Scientific papers cited in climate change mitigation and adaptation, high-carbon and ‘grey’ technology patents, by journal fields, 2015-20
Copy link to Figure 2.37. Scientific papers cited in climate change mitigation and adaptation, high-carbon and ‘grey’ technology patents, by journal fields, 2015-20

Percentage of articles cited by patents by ASJC journal fields

Figure 2.38. Share of scientific publications cited in climate change mitigation and adaptation patents by author country, 2006-13
Copy link to Figure 2.38. Share of scientific publications cited in climate change mitigation and adaptation patents by author country, 2006-13

As a percentage of the total publications, based on whole counts

Business R&D investment related to energy and the environment

Copy link to Business R&D investment related to energy and the environment

Measurement rationale

The previous sections, through their analysis of inventive activity and the R&I workforce, have touched on the contribution of businesses to generating knowledge related to achieving the energy and environmental SDGs. In market and mixed economies, businesses are key actors in R&I systems, developing and implementing new ideas to pursue economic opportunities. The business sector's role as a driver of total R&D expenditure is becoming increasingly important qualitatively and quantitatively (OECD, 2023[47]). However, there is considerable uncertainty as to how much of this growth has the potential to contribute to energy and environmental goals.

Policymakers are equally interested in ascertaining whether business R&D efforts result in incremental improvements in existing technologies or drive breakthroughs that potentially allow for new industries to develop. It is often presumed that, compared to R&D performed by government or non-profit institutions, business R&D is more results-driven, with a strong focus on immediate application, profitability and shareholder value. Such a statement may not apply equally to all types of business. To assess these issues, it is important to look not only at inventive outcomes but also to assess the resources businesses dedicate to such goals and how different types of businesses differ in their approaches.

Box 2.9. Measuring R&D relevant to energy and environmental goals in business surveys

Copy link to Box 2.9. Measuring R&D relevant to energy and environmental goals in business surveys

The main aggregate statistic used to describe R&D performance within the business enterprise sector is business enterprise R&D (BERD), the component of gross domestic expenditure on R&D (GERD) incurred by units in the business enterprise sector. Several countries measure R&D as relevant to energy and environmental goals but do so differently, preventing consolidated international comparisons. In their official business R&D surveys, few countries (e.g. Portugal) currently ask respondents to distribute their R&D expenditure across socio-economic objective (SEO) categories (see Chapter 4), such as energy and the environment. In addition to the need to keep surveys short, one key reason is that businesses do not necessarily view their R&D activities in terms of a classification system conceived for allocating government expenditures and funding.

OECD guidance in the Frascati Manual recognises that countries can find it useful to obtain information on the relevance of business sector R&D towards specific societal goals, but given thematic overlaps and diversity of priorities, suggest that this is not done on a mutually exclusive basis but through ad hoc questions (OECD, 2015[48]; Galindo-Rueda and López-Bassols, 2022[49]). Although there is no international reporting consensus at present, Box 2.10 presents examples of such practices for Canada and Norway. France’s business R&D survey collects information on R&D for three main resource and environmental goals but, similarly to SEO classifications, requests that the reported amounts be mutually exclusive.

Distributing R&D to economic activities (i.e. industries) is a possible alternative. However, it presents several challenges, starting from the point that economic activity is a rather ambiguous proxy for energy and environmental performance relevance. For example, a company in the R&D professional services sector may conduct R&D for an energy supplier company, or an equipment manufacturer may conduct R&D to improve the energy efficiency of its products or internal production processes. Industries are heterogeneous in their energy and environmental orientation for R&D and turnover. In none of these cases would the R&D expenditure be, based on the firm’s principal industrial activity, on the energy sector (“main activity” approach). Only the former case would be allocated to energy if a user or “industry-orientation” approach were followed. The Frascati Manual 2015 recommends classifying statistical units by their main activity and suggests that their R&D activities be classified both on that basis and by industry orientation.

Business R&D on energy and the environment

R&D in the energy, water and waste management utility industries

The proportion of business R&D companies conduct in the “electricity, gas and water supply; sewerage, waste management and remediation activities” is the highest in Latvia, Estonia and Portugal (Figure 2.39).

Figure 2.39. Business expenditure on R&D in the utilities and waste management industries, selected countries, 2012 and 2022
Copy link to Figure 2.39. Business expenditure on R&D in the utilities and waste management industries, selected countries, 2012 and 2022

Estonia and Lithuania experienced a substantial decline in the share of business R&D from 2012 to 2022. The United States, Romania and Israel dedicate the smallest share of business R&D to these industries. US (data expressed as interval) and French companies were the largest R&D performers in 2021 among countries for which data are available, including all major OECD countries but not China.

Box 2.10. Measurement case studies: Canada and Norway’s business R&D surveys

Copy link to Box 2.10. Measurement case studies: Canada and Norway’s business R&D surveys

Statistics Canada’s estimates of business R&D on clean technologies

Statistics Canada uniquely collects data on energy R&D expenditures disaggregated by area of technology as part of its Annual Survey of Research and Development in Canadian Industry. This approach follows IEA guidance, also reflected in OECD R&D measurement proposals, and allows for a detailed overview of the investments businesses make into relevant R&D. In 2022, businesses spent CAD1.9 billion (Canadian dollars) on in-house R&D to further develop “clean” (i.e. low-carbon) technologies. Energy efficiency saw the strongest growth in in-house R&D spending. This was followed by hydrogen and fuel cells, and electric power.

The percentage share of total in-house energy-related R&D for each of these three technology areas has also grown considerably since 2014. For instance, from 2014 to 2022, energy efficiency increased from 5.8% to 21.5%; hydrogen and fuel cells, from 3.6% to 11.1%; and electric power, from 3.7% to 13.5%. In-house R&D expenditures on fossil fuels also increased for the second consecutive year, up by CAD 171 million from 2021 to CAD 869 million (+24.5%) in 2022 (Figure 2.40).

Figure 2.40. In-house energy-related R&D expenditures on clean technologies, 2010-22
Copy link to Figure 2.40. In-house energy-related R&D expenditures on clean technologies, 2010-22

In CAD millions

Statistics Norway’s R&D in the business enterprise sector by thematic area

Statistics Norway similarly collects detailed data on R&D in the business sector, including by detailed technology thematic areas. The thematic areas are defined in sufficiently detailed terms to allow for the evaluation of trends with respect to business R&D in sustainable vs high-carbon energy, as well as environmental themes beyond energy. The petroleum thematic area saw a decline in business R&D in absolute terms between 2015 and 2022 by NOK 685 million (Norwegian kroner), while the sustainable energy and environmental sustainability-related thematic areas saw substantial increases. Businesses spent almost NOK 1.4 billion more on R&D in the energy efficiency and change theme (formerly energy efficiency), for instance, and almost NOK 1.5 billion more on environmental technology. The renewable energy and climate technology and other emission reductions thematic areas have also registered significant gains in business R&D spend.

Figure 2.41. Norway’s Business Expenditure on R&D by selected thematic area, 2015 and 2022
Copy link to Figure 2.41. Norway’s Business Expenditure on R&D by selected thematic area, 2015 and 2022

Current BERD cost (NOK millions)

R&D on energy

R&D in the business sector contributing to energy applications is significantly higher than R&D conducted by energy and other utilities, which account, on average, for less than 2% of the total. In the United States, where energy and environmental R&D questions have been used systematically for one decade, nearly 6% of business R&D is oriented to energy applications and 2% to environmental protection.

As shown in Figure 2.42, the energy utilities, extraction and petroleum and coal-product-transforming industries show the highest energy R&D intensity, i.e. percentage of energy-oriented R&D as a percentage of total industry’s R&D. However, these are not the largest contributors in absolute levels to energy R&D. Because of their much larger contribution to total R&D, the semiconductor and electronic component industry and the motor vehicle are the main investors in energy R&D, with shares close to 20% of total R&D.

New indicators of R&D investment by firms in selected markets and technology areas

As shown throughout this report, using existing data sources and classifications results in a reasonable but incomplete characterisation of the relevance of STI activities to energy and environmental goals, driving the need for new complementary data sources and indicators. This is also the case for business R&D, where the challenge stems from the heterogeneity of business activities and R&D portfolios and disclosure limits posed by the need to preserve assurances of confidentiality when collecting statistical data from firms.

The OECD Short-term Financial Tracker of Business R&D (SwiFTBeRD) dashboard helps monitor the latest trends in R&D investments by the world's top R&D business investors based on public annual and quarterly company reports (Box 2.11). For regulatory purposes, R&D disclosures are mandatory for large companies active in specific financial markets, and investors have come to expect regular reporting on those. While the motivation for exploring company reports stemmed originally from the limited timeliness of official statistics that results from annual or even biennial collection and reporting cycles, another set of possible insights from such data sources results from the qualitative information that is contained in company reports about the main lines of business of a company.

Figure 2.42. Business R&D expenditure in the energy application area across US industries, 2019
Copy link to Figure 2.42. Business R&D expenditure in the energy application area across US industries, 2019

Total expenditure and as a percentage of each industry’s R&D expenditure

Box 2.11. Measuring the latest business R&D trends using corporate reports

Copy link to Box 2.11. Measuring the latest business R&D trends using corporate reports

The OECD Short-term Financial Tracker of Business R&D (SwiFTBeRD) dashboard allows users to visualise quarterly, semi-annually and annually reported R&D data for the world's top R&D investors, providing company-specific and sectoral insights. It aims to deliver the timeliest possible view of R&D data reported by companies, with updates published continuously, shortly after they have been released in their quarterly financial reports. The initiative seeks to address the growing demand for real-time, high-quality R&D data to support evidence-based policy making, particularly in areas like innovation, economic growth and industrial strategy.

Company reports of R&D expenses need not coincide with R&D expenditures as covered in official R&D statistics compiled according to the Frascati Manual (OECD, 2015[53]). In order to compile the data presented in the SwiFTBeRD Dashboard, the OECD implements a series of adjustments to enhance comparability whenever the necessary information is available. Companies presenting their financial results in compliance with the International Financial Reporting Standards capitalise part of their development costs (under some criteria). In the data presented in SwiFTBeRD, capitalised development costs are added to reported R&D expenses, while amortisation of capitalised development expenditures is conversely excluded, provided that the information is available both in the annual and interim reports. In addition, when possible, expenses and impairment of purchased in-process R&D (as well as restructuring R&D costs) are excluded in the SwiFTBeRD figures so as to align as much as possible with R&D conducted in the reference period and deliver more meaningful indicators.

For the purposes of this publication, the OECD SwiFTBERD panel has been extended to include leading companies from the established alternative energy sector, which includes companies active in low-carbon energy technologies, such as Siemens Energy, Vestas Wind Systems, SolarEdge Technologies, Sungrow Power Supply, Orano, First Solar, SMA Solar Technology and Nordex. The electricity sector data include Electricité de France, Korea Electric Power, CGN Power, Iberdrola, GCL Technology, Energias de Portugal, ChargePoint, Landis+Gyr and Tokyo Electric Power.

Source: OECD (2025[54]), OECD SwiFTBeRD Dashboard, https://oecd-main.shinyapps.io/swiftberd/.

For the purposes of this publication and future monitoring, the OECD SwiFTBERD panel has been recently extended to include leading companies from the established alternative energy and electricity sectors. Analysis of trends for those companies shows robust growth in the R&D expenditure in this sector, with the growth since 2017 being second only to the software, computer and electronics sector (Figure 2.43). This performance, which is also matched by growth in revenue, needs to be considered in the context of ongoing supply chain difficulties faced by some of the companies in the group, as well as intensive competition from China, for which it is more difficult to obtain comparable data. The gap in R&D trends in real terms between alternative energy and electricity is increasing over time.

Figure 2.43. Trends in R&D and revenue for the SwiFTBERD panel of R&D investors, by industry groups, 2017-24
Copy link to Figure 2.43. Trends in R&D and revenue for the SwiFTBERD panel of R&D investors, by industry groups, 2017-24

References

[33] Aghion, P. et al. (2016), “Carbon Taxes, Path Dependency, and Directed Technical Change: Evidence from the Auto Industry”, Journal of Political Economy, Vol. 124/1, pp. 1-51, https://doi.org/10.1086/684581.

[14] Aristodemou, L. et al. (forthcoming), Assessing the directionality of R&D funding towards societal goals through new data sources and AI-assisted methods.

[24] Berger, M., A. Dechezleprêtre and D. Kirpichev Cherezov (forthcoming), Academic Start-ups’ Funding and the Role of Alternative Funding and Support Instruments, OECD Publishing, Paris.

[28] Bornmann, L. et al. (2022), “How relevant is climate change research for climate change policy? An empirical analysis based on Overton data”, PLOS ONE, Vol. 17/9, p. e0274693, https://doi.org/10.1371/journal.pone.0274693.

[8] Bush, V. (1945), Science: The Endless Frontier.

[35] Calel, R. and A. Dechezleprêtre (2016), “Environmental policy and directed technological change: Evidence from the European carbon market”, Review of Economics and Statistics, Vol. 98/1, pp. 173-191, https://doi.org/10.1162/rest_a_00470.

[10] Chapin, D., C. Fuller and G. Pearson (1954), “A New Silicon <i>p-n</i> Junction Photocell for Converting Solar Radiation into Electrical Power”, Journal of Applied Physics, Vol. 25/5, pp. 676-677, https://doi.org/10.1063/1.1721711.

[9] Dalla Fontana, S. and R. Nanda (2023), “Innovating to net zero: Can venture capital and start-ups play a meaningful role?”, Entrepreneurship and Innovation Policy and the Economy, Vol. 2, pp. 79-105, https://doi.org/10.1086/723236.

[34] Dechezleprêtre, A., R. Martin and M. Mohen (2013), Knowledge Spillovers from Clean and Dirty Technologies: A Patent Citation Analysis.

[30] Dernis, H. et al. (2015), World Corporate Top R&D Investors: Innovation and IP bundles, https://doi.org/10.2791/741349.

[6] European Commission (2024), Science, research and innovation performance of the EU, 2024 – A competitive Europe for a sustainable future, https://data.europa.eu/doi/10.2777/965670.

[21] Eurostat (n.d.), Students enrolled in tertiary education by education level, programme orientation, sex and field of education, https://ec.europa.eu/eurostat/databrowser/view/educ_uoe_enrt03__custom_15308082/default/table?lang=en&page=time:2021 (accessed on 11 March 2025).

[49] Galindo-Rueda, F. and V. López-Bassols (2022), “Implementing the OECD Frascati Manual: Proposed reference items for business R&D surveys”, OECD Science, Technology and Industry Working Papers, No. 2022/03, OECD Publishing, Paris, https://doi.org/10.1787/d686818d-en.

[22] Government of France (2024), Thèses soutenues en France depuis 1985, https://www.data.gouv.fr/fr/datasets/theses-soutenues-en-france-depuis-1985/ (accessed on 11 March 2025).

[23] Government of the United Kingdom (2023), “Insights from the UK-wide survey of the Research and Innovation Workforce 2022”, BEIS/DSIT Research Paper, No. 2023/004, https://assets.publishing.service.gov.uk/media/641d90305155a2000c6ad5f8/insights-uk-survey-research-innovation-workforce-2022.pdf.

[11] Green, M. (2000), “Photovoltaics: technology overview”, Energy Policy, Vol. 28/14, pp. 989-998, https://doi.org/10.1016/s0301-4215(00)00086-0.

[40] Guillard, C. et al. (2021), Efficient industrial policy for innovation: standing on the shoulders of hidden giants, CEP Discussion Papers, https://eprints.lse.ac.uk/113832/.

[26] Hallegatte, S. et al. (2016), “Mapping the climate change challenge”, Nature Climate Change 2016 6:7, Vol. 6/7, pp. 663-668, https://doi.org/10.1038/nclimate3057.

[31] Haščič, I. and M. Migotto (2015), “Measuring environmental innovation using patent data”, OECD Environment Working Papers, No. 89, OECD Publishing, Paris, https://doi.org/10.1787/5js009kf48xw-en.

[32] IEA/OECD (2021), Methodology for identifying fossil fuel supply related technologies in patent data, https://iea.blob.core.windows.net/assets/d14427c6-2aa2-4422-9074-5a68940a5a96/Methodology_for_identifying_fossil_fuel_supply_related_technologies_in_patent_data.pdf.

[15] IFC (2025), Innovation in Green Technologies: Opprotunity for Businesses in Low and Middle-income Countries?, https://www.ifc.org/content/dam/ifc/doc/2025/innovation-in-green-tech.pdf.

[12] IPCC (2023), Climate Change 2021 – The Physical Science Basis, Cambridge University Press, https://doi.org/10.1017/9781009157896.

[39] Martin, R. and D. Verhoeven (2022), The Private Value of Clean Energy Innovation.

[45] Marx, M. and A. Fuegi (2021), “Reliance on science by inventors: Hybrid extraction of in‐text patent‐to‐article citations”, Journal of Economics and Management Strategy, Vol. 31/2, pp. 369-392, https://doi.org/10.1111/jems.12455.

[44] Marx, M. and A. Fuegi (2020), “Reliance on science: Worldwide front‐page patent citations to scientific articles”, Strategic Management Journal, Vol. 41/9, pp. 1572-1594, https://doi.org/10.1002/smj.3145.

[13] Ministry of Higher Education and Science (2022), Rumbaseret grøn forskning – Bibliometrisk analyse af Danmarks rumbaserede grønne forskning September 2022, http://www.ufm.dk.

[54] OECD (2025), OECD SwiFTBeRD Dashboard, https://oecd-main.shinyapps.io/swiftberd/.

[17] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.

[18] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.

[47] OECD (2023), “The Impact of R&D tax incentives: Results from the OECD microBeRD+ project”, OECD Science, Technology and Industry Policy Papers, No. 159, OECD Publishing, Paris, https://doi.org/10.1787/1937ac6b-en.

[7] OECD (2021), OECD International Survey of Science, http://oe.cd/issa.

[16] OECD (2017), OECD Science, Technology and Industry Scoreboard 2017: The Digital Transformation, OECD Publishing, Paris, https://doi.org/10.1787/9789264268821-en.

[53] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris, https://doi.org/10.1787/978926.

[48] OECD (2015), Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris, https://doi.org/10.1787/9789264239012-en.

[29] OECD (2009), Patent Statistics Manual, OECD Publishing, Paris, https://doi.org/10.1787/9789264056442-en.

[50] OECD (n.d.), ANBERD Database, http://oe.cd/anberd.

[20] OECD (n.d.), Research and Innovation Careers Observatory, https://www.oecd.org/en/networks/research-and-innovation-careers-observatory.html.

[5] OECD (n.d.), Science and bibliometric indicators, https://www.oecd.org/en/data/datasets/science-and-bibliometric-indicators.html.

[37] OECD (n.d.), STI Micro-data Lab: Intellectual Property Database, http://oe.cd/ipstats.

[4] OECD and SCImago Research Group (CSIC) (2016), Compendium of Bibliometric Science Indicators, OECD, Paris, https://web-archive.oecd.org/2016-09-29/415063-Bibliometrics-Compendium.pdf.

[3] Perrons, R., A. Jaffe and T. Le (2020), “Tracing the linkages between scientific research and energy innovations: A comparison of clean and dirty technologies”, NBER Working Paper Series, http://www.nber.org/papers/w27777.

[46] Persoon, P., R. Bekkers and F. Alkemade (2020), “The science base of renewables”, Technological Forecasting and Social Change, Vol. 158, p. 120121, https://doi.org/10.1016/j.techfore.2020.120121.

[43] Poege, F. et al. (2019), “Science quality and the value of inventions”, Science Advances, Vol. 5/12, https://doi.org/10.1126/sciadv.aay7323.

[38] Popp, D. et al. (2020), “Innovation and entrepreneurship in the energy sector”, NBER Working Paper Series, No. 27145, https://doi.org/10.3386/W27145.

[36] Reeb, D. and W. Zhao (2020), “Patents do not measure innovation success”, Critical Finance Review, Vol. 9/1-2, pp. 157-199, https://doi.org/10.1561/104.00000087.

[25] Roche, M., A. Conti and F. Rothaermel (2020), “Different founders, different venture outcomes: A comparative analysis of academic and non-academic startups”, Research Policy, Vol. 49/10, p. 104062, https://doi.org/10.1016/j.respol.2020.104062.

[2] Rogelj, J. (2023), “Net zero targets in science and policy”, Environmental Research Letters, Vol. 18/2, https://doi.org/10.1088/1748-9326/acb4ae.

[41] Schmoch, U. (2008), Concept of a Technology Classification for Country Comparisons, http://193.5.93.80/edocs/mdocs/classifications/en/ipc_ce_41/ipc_ce_41_5-annex1.pdf.

[51] Statistics Canada (2024), Energy-related research and development expenditures, 2022, https://www150.statcan.gc.ca/n1/daily-quotidien/240919/dq240919b-eng.htm.

[52] Statistics Norway (n.d.), “11483: Thematic area of R&D in the business enterprise sector. Current cost, by detailed industry (SIC2007) (NOK million) 2015-2023”, Research and development in the business enterprise sector, https://www.ssb.no/en/statbank/table/11483/.

[27] Szomszor, M. and E. Adie (2022), “Overton: A bibliometric database of policy document citations”, Quantitative Science Studies, Vol. 3/3, pp. 624-650, https://doi.org/10.1162/qss_a_00204.

[19] UNESCO (2015), International Standard Classification of Education.

[42] van Raan, A. (2017), “Patent citations analysis and its value in research evaluation: A review and a new approach to map technology-relevant research”, Journal of Data and Information Science, Vol. 2/1, pp. 13-50, https://doi.org/10.1515/jdis-2017-0002.

[1] World Meteorological Organisation (2023), United in Science 2023, https://library.wmo.int/idurl/4/68235.

Notes

Copy link to Notes

← 1. The “planet” umbrella comprises: SDG 6: Clean water and sanitation; SDG 12: Responsible consumption and production; SDG 13: Climate action; SDG 14: Life below water; and SDG 15: Life on land.

← 2. High carbon patents include fossil fuels patents as identified in (IEA/OECD, 2021[32]) + patents in the following CPC - F02, C21B5, C21B7, C21B9, C10G1, C10L1, C10J, E02B, F01K, F02C, F22, F23, F27, F28, F24J, C01F7, C03B5, C04B7, C04B9, C04B11, C05C, C05D, C01C1, C10, C21B5, C21B7, C21B9, D21C3 ) but not among CPC Y02 patents

← 3. Grey technologies refer to fossil fuels patents as identified in (IEA/OECD, 2021[32]) + patents in the following CPC - F02, C21B5, C21B7, C21B9, C10G1, C10L1, C10J, E02B, F01K, F02C, F22, F23, F27, F28, F24J, C01F7, C03B5, C04B7, C04B9, C04B11, C05C, C05D, C01C1, C10, C21B5, C21B7, C21B9, D21C3 ) AND patents among CPC Y02 patents, plus patents in Y02E20, Y02T10/12, Y02T10/40, Y02P10/25, Y02P20/129, Y02P40/121, Y02P40/50, Y02P80/1

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment